Leaflet is an open-source JavaScript library for creating interactive maps. Unlike static visualization packages such as ggplot2 or ggmap, Leaflet maps are fully interactive and can include features such as:
Interactive panning/zooming
Pop-up tooltips and labels
Highlighting/selecting regions
It is used by many news organizations and tech websites to visualize geographic data. The leaflet package for R enables the creation of interactive maps within R without learning how to write JavaScript code. The leaflet documentation is a handy walkthrough for the basics of creating Leaflet maps in R. Let’s explore here how to create Leaflet maps using the same data we used to create raster maps with ggmap, crime data from the city of Chicago in 2017.1
Leaflet maps are built using layers, similar to ggplot2.
Create a map widget by calling leaflet()
Add layers to the map using one or more of the layer functions (e.g. addTiles(), addMarkers(), addPolygons())
Repeat step 2 as many times as necessary to incorporate the necessary information
Display the map widget
A basic example is:
m <-leaflet() %>%addTiles() %>%addMarkers(lng =-87.597241, lat =41.789829,
popup ="Saieh Hall of Economics")
m
0.2 Basemaps
Like ggmap, leaflet supports basemaps using map tiles. By default, OpenStreetMap tiles are used.
m <-leaflet() %>%setView(lng =-87.618994, lat =41.875619, zoom =12)
m %>%addTiles()
To use a different basemap provider, use addProviderTiles() with the name of the provider you wish to implement. You can view the complete list of providers using names(providers).
m %>%addProviderTiles(providers$Stamen.Toner)
m %>%addProviderTiles(providers$CartoDB.Positron)
m %>%addProviderTiles(providers$Esri.NatGeoWorldMap)
m %>%addProviderTiles(providers$Wikimedia)
0.3 Add markers
Markers are used to identify points on the map. Each point needs to be defined in terms of latitude/longitude coordinates. These can come from a variety of sources, most commonly either a map data file such as a shapefile or GeoJSON (imported using sf) or a data frame with latitude and longitude columns.
Let’s use the Chicago crimes data to draw a map of the city identifying the location of each reported homicide:
addMarkers() and related functions will automatically check data frames for columns called lng/long/longitude and lat/latitude (case-insensitively). If your coordinate columns have any other names, you need to explicitly identify them using the lng and lat arguments. Such as `addMarkers(lng = ~Longitude, lat = ~Latitude).
Without any customization, we get a basic map with each murder location indicated by a dropped pin. Each markers appearance can be customized, though the technical difficulty quickly ramps up. The awesome markers plugin offers the most straight-forward customizability options. Instead of using addMarkers(), use addAwesomeMarkers() to control the appearance of the markers using icons from the Font Awesome, Bootstrap Glyphicons, and Ion icons icon libraries. First you define the appearance of the icon using awesomeIcons(), then pass that as an argument to addAwesomeMarkers():
One concern is that some neighborhoods have so many murders that the points overlap. One solution enabled by Leaflet’s interactivity is to cluster markers at varying levels of detail using the clusterOptions argument to addMarkers():
Each point can have text added to it using either a label (appears either on hover or statically) or a popup (appears only on click). For instance we can label each murder with the date/timestamp when it was originally reported.
Leaflet can also draw spatial lines and shapes from R and add them to maps. Given our previous exposure to sf and importing shapefiles using st_read(), let draw a map of Chicago with each community area outlined.
areas <-here("static", "data",
"Boundaries - Community Areas (current)",
"geo_export_328cdcbf-33ba-4997-8ce8-90953c6fec19.shp") %>%st_read() %>%# convert community names to title casemutate(community =str_to_title(community))
## Reading layer `geo_export_328cdcbf-33ba-4997-8ce8-90953c6fec19' from data source `/Users/soltoffbc/Projects/Computing for Social Sciences/course-site/static/data/Boundaries - Community Areas (current)/geo_export_328cdcbf-33ba-4997-8ce8-90953c6fec19.shp' using driver `ESRI Shapefile'
## Simple feature collection with 77 features and 9 fields
## geometry type: MULTIPOLYGON
## dimension: XY
## bbox: xmin: -87.9 ymin: 41.6 xmax: -87.5 ymax: 42
## epsg (SRID): 4326
## proj4string: +proj=longlat +ellps=WGS84 +no_defs
To do this in ggplot(), we only need two lines of code:
The first several arguments adjust the appearance of each polygon region (e.g. color, opacity, border thickness). highlightOptions emphasizes the currently moused-over polygon. We can further add detail to this map by labeling each community area just as we did with points:
Now that we have a basic map of the city of Chicago with each community area identified, we can turn this map into a choropleth by filling in the color of each community area based on the number of reported homicides in 2017. First we calculate the total number of reported homicides by community area and merge this with the simple features data frame:
First, we define the bins. This is a numeric vector that defines the boundaries between intervals ((0,10], (10,20], and so on).
Then, we call colorBin() to generate a palette function that maps the RColorBrewer"YlOrRd" palette to our bins.
Finally, we modify addPolygons() to use the palette function and the density values to generate a vector of colors for fillColor(), and also add some other static style properties.
Let’s now modify the label to explicitly identify the number of reported homicides in the community area. We generate the HTML by hand and pass it to map(htmltools::HTML) so that Leaflet knows to treat each label as HTML instead of plain text. We also adjust the appearance of each label using the labelOptions argument and corresponding function.
